• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Hot XP 3200+

T

Toad21

Limp Gawd
Joined
Jun 29, 2004
Messages
445
So I got my water cooling setup today after about 10 hours of work on it. Man everything that could have gone wrong did. The GPU block I was going to use leaked, I am glad I water tested it.

So this cpu is in its own loop at running at 38c right now. WTF my air cooling ran at 40c.

I have:

Black Ice Extreme 120mm
Swiftech 6002-a
Hydro L30 pump (320g/h)

All on 1/2 ID tubing. Man this puppy runs hot. I was able to push it to 2.6ghz before the water heated up. The water is like room temperature to the touch yet the CPU reads 38c. This is on stock settings might I add. Not impressed at all. My Zalman 7000a-cu did almost as well.

UPDATE:

I now have it running at 33c. A CPU shim did the trick. My old heatsink had crushed the bottom pads, and the shim made the water block sit on the proc perfectly.
 
I'm not that surprised at those temps. my 2600 runs at 34 running stock voltage and speed. the main difference I see is in load temps. The extra thermal capacity of the water helps soak up the extra heat produced under load. a few other things to check would be:
mounting of the block - is it seated properly and is there enough pressure on the cpu die?
base of the block - is it rough or nicely lapped?
thermal compound - what type are you using if any? and has it been applied correctly?
radiator placement - is the rad getting cool air or pre warmed air, and is the airflow restricted at all?

other than that, is the watercooling quieter than your old aircooling? and what are your load temps like compared to what they were before?
 
DaMiEn said:
mounting of the block - is it seated properly and is there enough pressure on the cpu die?
base of the block - is it rough or nicely lapped?
thermal compound - what type are you using if any? and has it been applied correctly?
radiator placement - is the rad getting cool air or pre warmed air, and is the airflow restricted at all?

other than that, is the watercooling quieter than your old aircooling? and what are your load temps like compared to what they were before?
Click to expand...

I let me computer sit over night and now its 44c. Sigh.

Base of block had a mirror finish.

Arctic silver 5

Radiator is external WITH shroud so alot of air is going through radiator

It is not quiter. The temps are the same.

Sigh.
 
are you getting the temps from the mobo? if so then its because the temp probe on the mobo is(at least for me) under the cpu inside of the socket... so the reason it shows a high temps is because with air cooling there was air blowing on it...but with w ater cooling just the water is flowing over the cpu core ONLY....with air the air gets spread out. the water is concentrated on the cpu only
 
The temp is from the Motherboard Sensor Yes.

I am fairly certain the readings are right though, because when it got about 38c I could nolonger sustain 2.6ghz.

I can touch the water and it is ROOM temperature. If the proc is 40c, wouldnt that be pretty warm? Or is it the water isnt sooking up enough heat. The block is also room tempurature, what gives?

Thanks,

-Todd
 
Toad21 said:
The temp is from the Motherboard Sensor Yes.

I am fairly certain the readings are right though, because when it got about 38c I could nolonger sustain 2.6ghz.

I can touch the water and it is ROOM temperature. If the proc is 40c, wouldnt that be pretty warm? Or is it the water isnt sooking up enough heat. The block is also room tempurature, what gives?

Thanks,

-Todd
Click to expand...


When was the last time you had your finger calibrated? I doubt you can measure water temps accurately by touching it, so you need to use a thermometer. Measure water temp, ambient air temp, and try to get air temp coming out of the radiator. We can then probably tell you where your problem lies with regards to heat transfer.
 
I tapped my thermal diode to the CPU block and it reads a steady 84.6 farenheit. That is on top of the block.
 
My theories:

1. CPU is not transfering heat to block because of poor contact with cpu/thermal adhesive not used correctly.

-It was 9 hours in and 3am in the morning, and I remember putting thermal adhesive on cpu correctly(fairly thin coat only on center of cpu).

-It is a Swiftech 6002-a and the directions were not clear as to how far to loosen the screws. It says to loosen until even with the ridge or something. What ridge?

-I turned the pc on with the CPU block mounted to see if it would boot up. It would run for about 5 seconds then just shut off. I took the block off and tried without the block on and it would go for 1 second without shutting off. So the machine will shut off when the CPU gets too hot, should it have gotten that hot that quickly if the 1 pound Copper water block is on it?

2. The reservior is 1st in the loop. I feel the strength of the water coming out of the pump, and dont feel much pressure at all coming out of the reservior to the CPU.

Loop: Pump>Reservior>CPU>Radiator>Pump

Would placing the pump outlet directly on the cpu help some? Thats placing the CPU 2nd in the loop.

3. Radiator has a good amount of exhaust, and the water appears to be room temp, maybe a few Degress warmer. 40c would be 104 farenheit, and the water is Deffinetly colder than my skin.



Input greatly appreciated.
 
zer0signal667 said:
When was the last time you had your finger calibrated? I doubt you can measure water temps accurately by touching it, so you need to use a thermometer. Measure water temp, ambient air temp, and try to get air temp coming out of the radiator. We can then probably tell you where your problem lies with regards to heat transfer.
Click to expand...


Input from you is needed before it can be provided in return.
 
86.3 Farenheit water at the reservior. I can't run a temp test on the air on the outlet of the radiator, but it feels cold and with good airflow.
 
Is there an inlet and outlet for the Radiator? Is it possible to have it backwards?

Thanks.

-Todd
 
zer0signal667 said:
When was the last time you had your finger calibrated? I doubt you can measure water temps accurately by touching it, so you need to use a thermometer. Measure water temp, ambient air temp, and try to get air temp coming out of the radiator. We can then probably tell you where your problem lies with regards to heat transfer.
Click to expand...

I calibrated mine the last time my gf came over..err nm.
 
Toad21 said:
Loop: Pump>Reservior>CPU>Radiator>Pump

Input greatly appreciated.
Click to expand...

Res>x>x>x is optimal rather than having it second in your loop. Also, with water cooling you won't see much cooler idle temps than a high end air cooler. The difference becomes pronounced when you raise the voltage + OC a cpu/gpu. Under those kind of circumstances you could see a 20C+ rise in temperature using air but only a 5-6C rise using water.
 
Toad21 said:
86.3 Farenheit water at the reservior. I can't run a temp test on the air on the outlet of the radiator, but it feels cold and with good airflow.
Click to expand...


That seems sort of hot... but it's still not much of an indicator unless we know your ambient air temps. Assuming your pc is not in a sauna, high water temps would indicate that there is a problem moving heat from the water to the air. Do you have the fan pulling through or pushing into the radiator?
 
zer0signal667 said:
That seems sort of hot... but it's still not much of an indicator unless we know your ambient air temps. Assuming your pc is not in a sauna, high water temps would indicate that there is a problem moving heat from the water to the air. Do you have the fan pulling through or pushing into the radiator?
Click to expand...

He probably could use a dual 120mm radiator more than anything else.
 
The water is 86.4 farenhiet, The air is 70 farenhiet, and the CPU is 100 farenheat. The CPU is not transfering the heat to the water is my speculation.

Does that help? And noone has answered my question: Does it matter which barb is inlet or outlet on radiator? I have it going through on the left side, most setups I see go through on the right.
 
Toad21 said:
The water is 86.4 farenhiet, The air is 70 farenhiet, and the CPU is 100 farenheat. The CPU is not transfering the heat to the water is my speculation.

Does that help? And noone has answered my question: Does it matter which barb is inlet or outlet on radiator? I have it going through on the left side, most setups I see go through on the right.
Click to expand...

No it doesn't matter which barb is the inlet our outlet. Dude I think you just need a better/bigger rad.
 
this is ur problem: having the pump go to the reservoir, you aren't really getting ANY flow whatsoever to anywhere but ur res...redo the loop to go like this:

Res > Pump > CPU > Rad > Pump

that should do it!
 
Alex41290 said:
this is ur problem: having the pump go to the reservoir, you aren't really getting ANY flow whatsoever to anywhere but ur res...redo the loop to go like this:

Res > Pump > CPU > Rad > Pump

that should do it!
Click to expand...


No, that's not the problem. Do you think that more water goes into the res than comes out of it? Flowrate is constant throughout the system.

I would guess that a bigger radiator would solve the problem. If you're worried about transfer from the CPU to the water, then try remounting your block, making sure you have even pressure and complete contact.
 
Answer me this.

1. Is it normal for the Proc temp to be 15 degrees hoter than the water temp? I think that in lies my problem.

I just realized my CPU robber mounts were somewhat crushed on the bottom. Therefore I am ordering a CPU shim and going to see if that will help it make better contact.
 
Toad21 said:
Answer me this.

1. Is it normal for the Proc temp to be 15 degrees hoter than the water temp? I think that in lies my problem.

I just realized my CPU robber mounts were somewhat crushed on the bottom. Therefore I am ordering a CPU shim and going to see if that will help it make better contact.
Click to expand...

A 15C difference is not outrageous. And a shim will not help you get better contact, in fact it's more likely (although not very likely) that it will worsen it. The pads on the CPU are soft, they're not going to do much to keep the block from seating properly.
 
zero, what i meant was that you won't really get a "usable" flow going pump>res>etc.....anyway, the point is, try switching it around, i think it should make at least SOME difference, if not a big one
 
Alex41290 said:
zero, what i meant was that you won't really get a "usable" flow going pump>res>etc.....anyway, the point is, try switching it around, i think it should make at least SOME difference, if not a big one
Click to expand...


Please provide some evidence for this reasoning, because I don't really understand how it would make a difference.

Tie a 5 lb weight and a 10 lb weight to a rope at different lengths. Put the 10lb weight at the end and the 5 in the middle, then lift. Switch the positions of the weights, then lift again. Does it take more or less effort to life the rope the second time with the weights in a different order?
 
well, no matter how wrong i am, filling and bleeding must be a bitch to do...but i see your point...sorry for "talkin outta my ass"

EDIT: no thinking occured here whatsoever...this is what i was thinking: if you took just a pump and res and fill/bleed, wouldn't the res either overflow and drain to no flow to the pump, or wouldn't it just carry air throughout the system...but then i thought, the pump is also almost "sucking" the water through because of the propellor....so again, my bad for not thinking, and thanks for proving me wrong :D
 
The pads are crushed though, so it is NOT making a perfectly flat contact. Other than a CPU shim, what can i do?

Can I rebuy your pads and put new ones on?
 
Alex41290 said:
well, no matter how wrong i am, filling and bleeding must be a bitch to do...but i see your point...sorry for "talkin outta my ass"

EDIT: no thinking occured here whatsoever...this is what i was thinking: if you took just a pump and res and fill/bleed, wouldn't the res either overflow and drain to no flow to the pump, or wouldn't it just carry air throughout the system...but then i thought, the pump is also almost "sucking" the water through because of the propellor....so again, my bad for not thinking, and thanks for proving me wrong :D
Click to expand...


No need for apologies- I wasn't trying to prove you wrong, only show you a different way of looking at things. But you're right, res position can have a big impact on how easy filling/bleeding is.

Toad21 - you don't really need the pads there, I would just take them off. Their purpose is to keep the heatsink from mounting very unevenly. They may be able to help prevent you from rocking the heatsink so much that you break the core, but they're too spongy to ensure that the heatsink is actually mounted flat. I'm not familiar with that block offhand, but what you need to do is make sure that whatever mounting apparatus is used is being used properly (tghtened evenly, etc) and that there is nothing keeping the block from seating. Is there a raised ridge machined into the base to accomodate the socket?

I still think you have more room for improvement by means of radiator upgrade, but you might as well try this first.
 
OK. This is odd. I have the CPU shim coming in today or tomorrow, and i will post my results of that.

But for now, the Temp is 35c cpu core. ?? Why did it drop 5c, the room temp is the same, I havent done anything. I have heard of burning in the arctic silver or something, someone please explain.
 
Ok now my machine is running at a cool 33c idle, 35 load. It was the CPU shim that made all the difference. The pads that support the CPU block were crushed and or moved alittle on the bottom half of the proc, and it was sitting just alittle off center. With the CPU shim installed, it fit 100% level and I am very happy with it now.

Thanks for the help.

-Todd

P.S. I also put black tape around the fan where it was not air tight. That appeared to make a 1 or 2c difference as well, but only after I had the shim installed. Next I am going to put a another fan on the radiator as a puller. I will let you know what my results are.
 
You must log in or register to reply here.
Back
Top